Michael D. Smith is CEO of CPower.
PJM Interconnection’s record-high capacity prices for next year may send a build signal to generators, but the nation’s largest grid operator needs more supply to meet accelerating demand growth now — and virtual power plants can help.
A cheaper and faster-to-market alternative than building generation, VPPs use existing distributed energy resources, including batteries, EVs, smart thermostats and other controllable devices and loads, to keep the lights on when the grid is stressed. In aggregating customer DERs and dispatching them together as VPPs, we can quickly and affordably balance the grid today, without the lengthy waits and large investments needed to build physical power plants.
Costly problem
PJM attributed its sharp increase in capacity prices — from $28.92/MW-day for 2024/2025 to $269.92/MW-day for 2025/26 — to generator retirements, increased electricity demand and market reforms, while some market observers put the onus on the grid operator. Advanced Energy United, for one, blamed poor planning and insufficient electric transmission build-out by PJM.
Whatever the cause, the challenge is real. Not only are generation resources retiring faster than their capacity is being replaced, but the projected peak load is rising, due largely to AI, data centers and EVs. In addition, the market is adapting to Federal Energy Regulatory Commission-approved regulations for improved reliability risk modeling for extreme weather and accreditation that more accurately values each resource’s contribution to reliability.
With multiple factors in play and no way to bring enough new generation online in the short term to ensure grid reliability in PJM’s next delivery year, we have little choice but to consider alternatives such as VPPs, which we can quickly stand up without building physical infrastructure, or which are located on the distribution system and can be connected sooner than through the PJM process.
Proven solution
While PJM has not avoided VPPs, it could do more with them, beginning by using demand response more than it does now. A key component in VPPs, demand response accounted for only 5% of the capacity that PJM has secured for next year, despite its dependability and history of improving grid reliability.
After all, it has not yet been two years since VPPs including demand response helped keep the lights on in PJM amidst widespread failures in its generation fleet. During Winter Storm Elliott in December 2022, reducing usage and keeping load down was key to avoiding blackouts. Almost a quarter of the grid operator’s generation capacity — 47,000 MW — was on forced outages at one point when temperatures plunged during the Christmas weekend.
Long a linchpin of grid operators’ handling of peak summer loads, demand response has become a year-round solution as extreme weather events like Winter Storm Elliott have become more common and severe. With a history of handling summer heatwaves as well as winter storms, demand response is critical to ensuring a reliable grid, according to a report about grid reliability challenges co-published by Wood Mackenzie and CPower. Thus, demand response should play a greater role in PJM’s capacity markets.
Energy efficiency
PJM should similarly embrace energy efficiency, contrary to recent efforts to end its participation in the grid operator’s capacity market. Like demand response, energy efficiency is an essential part of VPPs.
The grid operator’s independent market monitor, or IMM, filed a FERC complaint in July regarding energy efficiency in PJM. The IMM wrongly argued that “PJM has been paying the capacity market clearing price to EE resources that do not meet the definition of an EE Resource,” and therefore that PJM should stop payments to resources that allegedly do not meet its criteria for energy efficiency. Further, IMM urges that EE providers (and by implication, customers) should be forced to return past payments that PJM previously approved.
Not only is the IMM’s argument contrary to PJM’s approved tariff, but it is also wrong economically. PJM’s sharply higher capacity prices for 2025/26 clearly demonstrate the need for more diversification of resources in the market, not less.
Reliable future
Even though PJM’s capacity prices hit record highs for 2025/26, they could go even higher or remain elevated unless the grid operator’s generation capacity can catch up with its surging demand.
The 8.9% decline in total capacity offered by non-energy efficiency resources in PJM’s latest capacity auction, from 148,945.7 MW in 2024/2025 to 135,692.3 MW in 2025/2026, marked the fourth consecutive year-over-year drop. Meanwhile, the peak load forecast increased by 2.2% to 153,883 MW, and the grid operator’s overall reserve margin shrunk by almost two full percentage points, from 20.4% to 18.5%.
With a total cost to load of $14.7 billion for the next delivery year due to tighter supply relative to demand, customers in PJM can ill afford to be stuck with another big bill for capacity. Yet, they may well face a double whammy with the 2026/27 capacity auction set for December. Exelon has already warned of double-digit rate increases for some of its utility subsidiaries due to the results for 2025/26.
PJM has stated that much of the struggle to bring new generation online is beyond its control, noting that 38,000 MW of resources have cleared its interconnection queue but have not been built due to external challenges, including financing, supply chain and siting/permitting issues. Thus, even if PJM has more generation capacity in 2026 than in 2025, it will not be enough to bring prices down dramatically.
So, with great time and expense needed, even if generators react to the build signal sent by the recently released capacity prices for 2025/26, not enough will be built anytime soon. Therefore, the best solution, and perhaps the only one, is to quickly scale up VPPs that include demand response and energy efficiency resources.
We may need more power generation resources, but nothing says those resources must be large, central station grid-scale power plants. Sometimes smaller and more distributed is better.
